EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Cleaning codes for speculative attention mode (#10149)

Browse Source
This commit is contained in:
Baizhou Zhang
2025-09-08 17:38:06 -07:00
committed by GitHub
parent 148022fc36
commit 8ad700f735
7 changed files with 14 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/speculative/eagle_worker.py
View File

@@ -191,7 +191,7 @@ class EAGLEWorker(TpModelWorker):
# Initialize decode attention backend
self.draft_attn_backend = self._create_decode_backend()
# Initialize draft extend attention backend (respects speculative_attention_backend setting)
# Initialize draft extend attention backend (respects speculative_attention_mode setting)
self.draft_extend_attn_backend = self._create_draft_extend_backend()
self.draft_model_runner.draft_attn_backend = self.draft_attn_backend
@@ -236,7 +236,7 @@ class EAGLEWorker(TpModelWorker):
}
backend_name = (
"decode_attention_backend"
if self.server_args.speculative_attention_backend == "decode"
if self.server_args.speculative_attention_mode == "decode"
else "prefill_attention_backend"
)
return self._create_backend(